基于贝叶斯最大后验估计的个人网上银行交易风险管理
Risk Identification of Personal Online Banking Transactions Based on Bayesian Maximum A Posteriori Estimation

以某全国性股份制银行省级分行辖内的零售规模相近的支行网点为样本，基于历史个人网上银行交易额和问卷数据，运用贝叶斯最大后验估计方法，推断未来交易额的合理区间，对相关潜在风险进行预警和识别。结果显示：个人网上银行交易风险多源于钓鱼网站诱使个人账户资金频繁转出，外部欺诈事件在节假日期间多发，源自节假日期间由网络购物导致的电子银行转账交易量的大幅增加，商业银行在节日较多的月份须更加关注网上银行的交易风险。
In this paper, a sample of branch outlets of similar retail-banking scale of a shareholding bank in Hunan Province is used to estimate the personal online banking transaction data based on Bayesian maximum a posteriori estimation method. The empirical results indicate that personal online banking transactions in months with more holidays are higher, which may result from more frequent online shopping of people in those months. Meanwhile, phishing sites become more active, inducing more account funds out through the Internet Bank channels, so there are more personal online banking transactions. Therefore, the commercial banks should be more on guard against the risk of online banking transactions in months with more holidays.